Scholarly article on topic 'Mass Transfer Characteristics in a Standard Pulsed Sieve-plate Extraction Column'

Mass Transfer Characteristics in a Standard Pulsed Sieve-plate Extraction Column Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
OECD Field of science
Keywords
{"pulsed sieve-plate column" / "overall volumetric mass transfer coefficient" / "axial dispersion model" / "axial mixing" / "steady- state concentration profiles"}

Abstract of research paper on Chemical engineering, author of scientific article — Caishan Jiao, Shuai Ma, Qiong Song

Abstract Pulsed sieve-plate extraction column is the very important equipment in nuclear spent fuel reprocessing. Study of the mass transfer characteristics in the extraction column under different operating conditions is beneficial to the optimization and scale-up of the columns. The experiments were carried out in the system of 30% tributyl phosphate (TBP) in kerosene-nitric acid-deionized water in the standard pulsed sieve-plate column. The axial dispersion model (ADM) coupled with the steady-state concentration profiles was used to calculate the overall volumetric mass transfer coefficient K ox a. The influence of the continuous and dispersed phase superficial velocities, pulse intensity and flow rate on the K ox a had been analyzed. It shows that with the increase of the continuous and dispersed phase superficial velocities and flow ratio, K ox a increases first and then decreases; but it increases when the pulse intensity increases. Finally, an empirical correlation had been concluded to predict the mass transfer performance in the extraction column.

Academic research paper on topic "Mass Transfer Characteristics in a Standard Pulsed Sieve-plate Extraction Column"

Available online at www.sciencedirect.com

ScienceDirect

Energy Procedia 39 (2013) 348 - 357

Asian Nuclear Prospects 2012 (ANUP2012)

Mass Transfer Characteristics in a Standard Pulsed Sieve-Plate Extraction Column

Jiao Caishana,Ma Shuaia,*,Song Qionga

_aCollege of Nuclear Science and Technology,Harbin Engineering University,Harbin 150001,China_

Abstract

Pulsed sieve-plate extraction column is the very important equipment in nuclear spent fuel reprocessing. Study of the mass transfer characteristics in the extraction column under different operating conditions is beneficial to the optimization and scale-up of the columns. The experiments were carried out in the system of 30% tributyl phosphate (TBP) in kerosene-nitric acid-deionized water in the standard pulsed sieve-plate column. The axial dispersion model (ADM) coupled with the steady-state concentration profiles was used to calculate the overall volumetric mass transfer coefficient Koxa. The influence of the continuous and dispersed phase superficial velocities, pulse intensity and flow rate on the Koxa had been analyzed. It shows that with the increase of the continuous and dispersed phase superficial velocities and flow ratio, Koxa increases first and then decreases; but it increases when the pulse intensity increases. Finally, an empirical correlation had been concluded to predict the mass transfer performance in the extraction column.

© 2013 The Authors. Published by Elsevier Ltd.

Selection and peer-review under responsibility of Institute of Nuclear and New Energy Technology, Tsinghua University

Keywords:pulsed sieve-plate column; overall volumetric mass transfer coefficient; axial dispersion model; axial mixing; steady-state concentration profiles;

* Corresponding author. Tel.: +1-524-507-1679; fax: +0-451-8251-8400. E-mail address: s310150068@hrbeu.edu.cn.

1876-6102 © 2013 The Authors. Published by Elsevier Ltd.

Selection and peer-review under responsibility of Institute of Nuclear and New Energy Technology, Tsinghua University doi: 10. 1016/j .egypro .2013.07.222

1. Introduction

With the development of the nuclear power industry, nuclear reprocessing has been playing an increasing role in the nuclear fuel cycle. For the time being, there are many processes which have been proposed in nuclear reprocessing. PUREX (Plutonium Uranium Reduction Extraction) is one of the most well-known aqueous processes in nuclear industry. So nowadays, pulsed sieve-plate columns based on PUREX have been widely used to extract some useful nuclides from spent nuclear fuel, e.g. uranium and plutonium.

Pulsed sieve-plate column provides high energy to break droplets which assists to attain large interfacial area to increase the mass transfer efficiency'11. Due to the separation from the working zone with highly radioactive waste, the pulse generator is easily maintained. Therefore, pulsed sieve-plate column has a clearer advantage than other extraction columns for having no internal moving parts. Pulsed sieve-plate column provides more uniform distribution phases across the column cross section with less tendency toward flow channeling which causes the easier scale-up as compared to columns with rotatory agitators'1' 2]. On account of these advantages, pulsed sieve-plate columns develop a good extension in nuclear reprocessing plant.

Due to the complex behavior of the hydrodynamics and mass transfer performance, the research concerning the design and scale-up of extraction columns is still far from satisfactory, although a great deal of experimental efforts has been expended'31. The introduction of the mechanical agitator or pulsation in extraction equipment often enhances the interface turbulence and generates a large interfacial area per unit volume of the fluid mixture. Consequently, it increases axial mixing in the phases and lowers the driving force available for the mass transfer'41. The plug-flow computational model without taking account of the axial mixing leads to an overestimation of the true value. It has been reported that axial mixing can decrease the mass transfer efficiency as much as 30%'5]. Thus, an axial dispersion model which has been verified well'6 7] was adopted to consider the deviation from the ideal plug-flow situation.

The axial dispersion model coupled with steady-state concentration profile method was used to evaluate the overall volumetric mass transfer coefficient, and the dispersed phase holdup was measured by the volumetric replacement method'81.

Nomenclature

a specific surface area, m2/m3

A pulse amplitude, cm

/ pulse frequency, s-1

e axial dispersion coefficient, cm2/s

h height of the sampling point

H effective column height, cm

Hox 'true' height of transfer unit

Ko^a overall mass transfer coefficient based on continuous phase, s"

Nox 'true' number of transfer unit

Pe Peclet number

L flow ratio

u superficial velocity, cm/s

x continuous phase concentration, mol/L

y dispersed phase concentration, mol/L

Z dimensionless height

Greek symbols p density, kg/m3 H viscosity, mPa-s y interfacial tension, N/m </> dispersed phase holdup, %

Subscripts

E extract liquid

F feed

R raffinate

S solution

x continuous phase

y dispersed phase

Superscripts

equilibrium

2. Experimental

2.1. Apparatus

The experiment was carried out in a 38 mm inner diameter extraction column with the total height of 2324 mm. Along the column, 30 sieve-plates were fixed on a vertical central rod, the free cross-sectional area fraction of each plate was 23%, the thickness of plate was 0.8 mm, and plate spacing was 50 mm. Each phase was fed by a peristaltic pump, and the phase superficial velocity was measured according to the linear correlation based on the pump rotating speed. The two phases flowed countercurrently through the column, while the pulse generator was operated for the phases stirring through an electronic motor. The schematic diagram of the experimental apparatus is shown in Fig 1.

aqueous stream organic stream air supply

Fig. 1. Schematic diagram of the experimental apparatus 1. feed tank 2. solvent tank 3. raffinate tank 4. extract liquid tank 5. peristaltic pump 6. sieve-plate 7. sample point 8. pulse leg 9. pulse generator

In order to get the samples of the two phases in the operating state, eight pair of sampling needles was inserted through the rubber stopper along the column. To sample the pure continuous phase, a small flared sleeve filled with hydrophilic polypropylene non-woven fabric was set at the end of a stainless needle which was inserted into the centre of the column just above the sieve plate facing up[7]. Similarly, a small flared sleeve filled with hydrophobic polypropylene fibre was set at the end of a stainless needle which was inserted just below the plate facing downwards. Through the sampling needles which had been modified, the pure samples of two phases could be taken simultaneously.

2.2. System

The system used in this experiment was 30% tributyl phosphate (TBP) in kerosene-nitric acid -deionized water. HNO3 solution was the solute, and the concentration was 3.00+0.02 mol/L. The direction of mass transfer was from continuous phase to dispersed phase while 30% TBP in kerosene was used as

the dispersed phase. The TBP and HNO3 were all analytically pure. The physical properties of these chemical substances are listed in Table 1. The equilibrium concentration relation of HNO3 in two phases at the room temperature was illustrated in Fig 2, the correlation was given as follows: x*=3.50y2+1.55y+0.20 (y > 0.06 mol/L) (1)

Table 1. Physical properties of the system at room temperature

Material p (kg/m3) H (mPa-s) y (N/m)

Continuous phase 1097.5 0.00124

0.02155

Dispersed phase 856.1 0.00274

y (mol/L)

Fig. 2. The equilibrium concentration curve of HNO3 in two phases

The operating variables systematically investigated in this experiment were continuous and dispersed phase superficial velocities, pulse intensity, and flow ratio. The range of variables investigated is shown in Table 2.

Table 2. Range of the operating variables

wxxl0-4(m/s) w„x10-4(m/s) Ap< 10-2(m/s) R

2.63~13.13 8.53~20.83 0.78~1.33 0.8~4

In each run, when the volume of the raffinate was about five or six times than that of the column and the HNO3 concentration of the continuous phase remained constant, the steady-state of mass transfer was achieved'91. The volume of the samples withdrawing from every sampling point was 0.5 to 1ml, which had little effect on the flow condition in the column. The concentration of all the samples was determined by an automatic titrator immediately after each experiment.

2.3. Mathematic modeling and algorithm

The axial dispersion model is a modification of the plug flow model without considering the unfavourable effect of the axial mixing. The axial dispersion coefficient is a key parameter which

quantifies the deviation from ideal plug flow. The factors introducing to the axial mixing are the presence

of circulatory flow, molecular diffusion, small eddies and channelling'11.

The seven assumptions below should be considered in the analysis of the mass transfer process based

on the axial diffusion model[10].

• The continuous and dispersed phase superficial velocities do not change with the height.

• The continuous and dispersed phases are immiscible (or their solubility does not vary with solute concentration and height).

• The mean velocity and concentration of each phase can be regard as identical at the same column cross section.

• The effect of the sampler needle on the phases flow is little and can be neglected.

• The distribution coefficient is constant, not a function of concentration.

• The product of the mass transfer coefficient and interfacial area per unit volume is constant throughout the column.

• The gradients of solute concentration in each phase are continuous.

~e*dz\z JL1

ZxdZlZ+dZ

uxc\z+dz

■zM\z

Koxrtx-x*)

~ZySZ\z+dZ uy\z+dz

UxXR uys

Fig. 3. Schematic diagram of the axial dispersion model

With reference to Fig 3, based on the solute mass balance and the axial dispersion model, the differential equation set for steady-state process was listed as follows:

1 d2 x dx

Pzx dz2 dz 1 d 2y dy

(x — x ) = 0

Pz„ dz2 dz

'Nox~ ( x — x ) = 0

where Pzx=

zx (1- x )

, Pzy= -

, z=h/H,

The inlet and outlet concentrations can be determined; therefore the boundary conditions are set below:

At Z=0, - dZ =Pex(xF-xz=o ) and - ~Z

At Z=1,

=0 and dZ =Pey (yZ=i'-yS)-

To solve the Eqs. (2) and (3), the forth order explicit Runge-Kutta method was applied to achieve the analytic solutions. Then the fminsearch iteration coupled with the measured solute concentration profiles was adopted to optimise the Pex, Pey and Nox through the Matlab software. When the Nox was achieved, the Koxa can be calculated from the formula 4 below.

3. Results and discussion

3.1. Effect of the continuous phase superficial velocity on Koxa

The influence of the continuous phase superficial velocity (ux) on Koxa is illustrated when uy and pulse intensity were kept constant. It has been observed that Koxa increases with the increase of ux until it reaches a maximum value. At first, the increase of ux makes the dispersed phase droplets distribute in continuous phase uniformly, therefore the continuous phase axial mixing decreases and mass transfer efficiency increases. When it reaches the maximum, the faster the continuous phase superficial velocity, the shorter the phases contact. And the aggravation of the turbulence induced by the higher continuous phase superficial velocity introduces a tendency of the increase of the continuous phase axial mixing, so

Fig. 4. Effect of ux on Koxa Fig. 5. Effect of uy on Koxa

3.2. Effect of the dispersed phase superficial velocity on Koxa

As the continuous phase superficial velocity and pulse intensity are constant, Koxa increases with the increase of the dispersed phase superficial velocity (uy) firstly, and then decreases, as shown in Fig 5. Generally, axial mixing of the continuous phase decreases with the increase of uy[8]. The increase of dispersed phase superficial velocity causes an increase in dispersed phase holdup, resulting in the rise of mass transfer interfacial area and driving force among the phases. The increasing of the dispersed phase superficial velocity will lead to a higher dispersed phase holdup, the dispersed phase produces large

droplets. As a result, the overall volumetric mass transfer coefficient decreases gradually with the decrease of the specific surface between the two phases.

3.3. Effect of the pulse intensity on KoXa

The effect of the pulse amplitude and frequency on Koxa can be seen as a unity of their product, i.e. the pulse intensity (A/). Illustrated in Fig 6, with the increase of the pulse intensity, the Koxa increases incipiently slowly, and then rapidly. Ebrahimi et al. [11] and Torab-Mostaedi and Safdari [12] has reported that the mass transfer performance increases with an increase of pulse intensity. The increase of pulse intensity can improve the mass transfer driving force greatly, and the specific surface area will increase with the decrease of the dispersed phase droplet diameter caused by the pulse effect. Therefore, Koxa increases. However, it should be note that higher pulse intensity will aggravate the axial mixing, leading to the decline of the mass transfer efficiency [13].

Mx=7.88x10"4 m/s

By=1.673x10"3 m/s

0.8 1.0 1.2 Af.< 10"2 (m/s)

Fig. 6. Effect of pulse intensity on Koxa

Fig. 7. Effect of flow ratio on Koxa

3.4. Effect of the flow ratio on Koxa

The flow ratio (L) has an important impact on the mass transfer characteristics. The overall volumetric mass transfer coefficient has been researched when the continuous phase superficial velocity and pulse intensity are constant. As shown in Fig 7, the overall volumetric mass transfer coefficient has a tendency to rise with the increase of the flow ratio firstly, and then reduces. The dispersed phase holdup and special superficial area increase simultaneously as the flow ratio increases, while decrease after the maximum. The increase of the special superficial area will improve the volumetric mass transfer coefficient. Hence the Koxa first increases and then decreases following with the change of the special superficial area.

3.5. The empirical correlation of Koxa

The experiments were carried out under different operating conditions. In the discussions above, Koxa varies with these three variables, i.e. the continuous and dispersed phase superficial velocities and pulse intensity. According to the empirical correlation proposed by Eguchi. K[14], Koxa can be predicted by the equation as follow:

Koxa=5.45Uxa5V71(Afia07 (5)

The calculated values of Koxa from the Eq. (5) were compared with the experimental data as shown in Fig 8 with a maximum deviation of 35% which was practically acceptable.

Fig. 8. Comparison of Koxaexp with Koxacai

4. Conclusions

A study of the mass transfer performance in the standard pulsed sieve-plate column was presented in this paper. The axial dispersion model coupled with the steady-state concentration profiles was applied to evaluate the actual flow characteristics well. From the experiments, it can be concluded that the overall volumetric mass transfer coefficient increases with the continuous and dispersed phase superficial velocities and flow ratio firstly, and then decreases; but it increases with the increase of the pulse intensity under the specific column dimensions and material system.

An empirical correlation was proposed to predict the Koxa which was in good agreement with the experimental values.

Further experimental efforts should be focused on different systems so as to predict the performance the design and scale-up of the pulsed sieve-plate column more accurately.

References

[1] Kolhe NS, Mirage YH, Patwardhan AV, et al. CFD and experimental studies of single phase axial dispersion coefficient in pulsed sieve plate column. Chemical Engineering Research and Design; 2011,89(10),1909-1918.

[2] Godfrey JC, Marrocchelli A, Marley ST, et al. Continuous phase axial mixing in pulsed sieve plate liquid-liquid extraction columns. Chemical Engineering Research and Design; 1988,66, 445-457.

[3] Luo GS, Li HB, Fei WY, et al. A Simplified Correlation of Mass Transfer in a Pulsed Sieve Plate Extraction Column. Chemical Engineering & Technology; 1998,21(10),823-827.

[4] Kannan A, Harikrishman TL, Varma YBG, et al. A numerical technique for the estimation of the true mass transfer coefficient in extraction columns. The Chemical Engineering Journal; 1990,45(2), 133-135.

[5] Li NN, Ziegler EN. Effect of axial mixing on mass transfer in extraction columns. Industrial & Engineering Chemistry;1967,59(3),30-36.

[6] Danckwerts PV. Continuous flow systems: Distribution of residence times. Chemical Engineering Science, 1953,2(1),1-13

[7] Smoot LD, Babb AL. Mass Transfer Studies in a Pulsed Extraction Column: Longitudinal Concentration Profiles. Industrial & Engineering Chemistry Fundamentals; 1962,1(2), 93-103.

[8] Tang XJ, Luo GS, Wang JD. A dynamic forward mixing model for evaluating the mass transfer performances of an extraction column. Chemical Engineering Science; 2004,59(21), 4457-4466.

[9] Wang YY, Jing S, Wu QL, et al. Axial mixing and mass transfer characteristics of pulsed extraction column with discs and doughnuts. Transactions of Nonferrous Metals Society of China; 2006,16(1),178-184.

[10] Sleicher CA. Axial mixing and extraction efficiency. AIChE Journal, 1959,5(2),145-149.

[11] Ebrahimi SSS, Bahmanyar H, Safdari J. A New Enhancement Factor for Mass Transfer Coefficients and Designing Pulsed Sieved-Plate Columns. Journal of Chemical Engineering of Japan; 2009,42(1),10-20.

[12] Torab-Mostaedi M, Safdari J. Prediction of mass transfer coefficients in a pulsed packed extraction column using effective diffusivity. Brazilian Journal of Chemical Engineering; 2009,26(4),685-694.

[13] Safari A, Safdari J, Abolghasemi H, et al. Axial mixing and mass transfer investigation in a pulsed packed liquid-liquid extraction column using plug flow and axial dispersion models. Chemical Engineering Research and Design; 2012,90(2),193-200.

[14] Eguchi W, Nagata S. Studies on the pulsed plate column. Chem. Eng. Japan, 1959,23(3),146.